JP2023096505A - Pedestrian traverse prediction method and pedestrian traverse prediction device - Google Patents

Abstract

To provide a pedestrian traverse prediction method and pedestrian traverse prediction device capable of accurately predicting whether pedestrians are to cross a road when the pedestrians exist on both sides of the road.SOLUTION: A pedestrian traverse prediction device includes an object detection device and controller mounted on an own vehicle 50. The object detection device detects a pedestrian 60 within a predetermined distance in front of the own vehicle 50. The object detection device detects a pedestrian 61 on the opposite side of a road with a position of the pedestrian 60 as reference. When a behavior change of the pedestrian 61 is detected within a predetermined time after the object detection device detects a behavior change of the pedestrian 60, the controller allows at least one likelihood between a first likelihood when the pedestrian 60 crosses the road and a second likelihood when the pedestrian 61 crosses the road to be higher than that when the behavior change of the pedestrian 61 is not detected within the predetermined time.SELECTED DRAWING: Figure 2

Description

The present invention relates to a pedestrian crossing prediction method and a pedestrian crossing prediction device.

Conventionally, when turning left or right at an intersection, there is known a method of predicting the possibility of intersection between the two based on the traveling direction and speed of the vehicle after the right or left turn and the traveling direction and speed of the pedestrian (Patent Reference 1).

JP 2014-232412 A

However, in the invention described in Patent Document 1, if there are pedestrians on both sides of the road, there is a possibility that one pedestrian will accelerate to cross the pedestrian crossing in response to a change in behavior of one pedestrian. It is not taken into consideration, and there is a risk that the prediction accuracy will decrease.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and its object is to provide a walking device capable of accurately predicting whether or not a pedestrian will cross the road when there are pedestrians on both sides of the road. It is to provide a pedestrian crossing prediction method and a pedestrian crossing prediction device.

A pedestrian crossing prediction method according to an aspect of the present invention detects a first pedestrian within a predetermined distance in front of a vehicle, and detects a second pedestrian on the other side of the road based on the position of the first pedestrian. If a change in behavior of the second pedestrian is detected within a predetermined time after detecting a change in behavior of the first pedestrian, or if a change in behavior of the second pedestrian is not detected within a predetermined time. At least one of the first likelihood that the first pedestrian crosses the road and the second likelihood that the second pedestrian crosses the road is increased.

Advantageous Effects of Invention According to the present invention, it is possible to accurately predict whether a pedestrian crosses the road when there are pedestrians on both sides of the road.

FIG. 1 is a configuration diagram of a pedestrian crossing prediction device according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a likelihood calculation method. FIG. 3 is a diagram illustrating an example of a likelihood calculation method. FIG. 4 is a diagram illustrating an example of a likelihood calculation method. FIG. 5 is a flowchart for explaining an operation example of the pedestrian crossing prediction device. FIG. 6 is a diagram illustrating an example of a likelihood calculation method. FIG. 7 is a diagram illustrating an example of a likelihood calculation method. FIG. 8 is a diagram illustrating an example of a likelihood calculation method.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the description of the drawings, the same parts are denoted by the same reference numerals, and the description thereof is omitted.

A configuration example of a pedestrian crossing prediction device will be described with reference to FIG. As shown in FIG. 1 , the pedestrian crossing prediction device includes an object detection device 1 , a vehicle position estimation device 2 , a map database 3 and a controller 100 .

A pedestrian crossing prediction device is mounted on a vehicle (self-vehicle). The pedestrian crossing prediction device may be installed in a vehicle having an automatic driving function, or may be installed in a vehicle without an automatic driving function. Also, the pedestrian crossing prediction device may be installed in a vehicle capable of switching between automatic driving and manual driving. In addition, even if the automatic driving function is a driving support function that automatically controls only a part of the vehicle control functions such as steering control, braking force control, and driving force control to assist the driver's driving. good.

The object detection device 1 includes object detection sensors such as a laser range finder, radar, lidar, camera, and sonar. The object detection device 1 detects objects around the own vehicle using a plurality of object detection sensors. The object detection device 1 detects moving objects such as other vehicles, buses, motorcycles, bicycles, and pedestrians, and stationary objects such as crosswalks and bus stops. The object detection device 1 detects the position, attitude, size, speed, acceleration, etc. of a moving object and a stationary object with respect to the own vehicle. The object detection device 1 outputs information regarding the detected object to the controller 100 . Since object detection methods using laser range finders, radar, lidar, cameras, sonar, etc. are well known, detailed description thereof will be omitted.

The own vehicle position estimation device 2 detects the position (position information) of the own vehicle on the ground. The vehicle position estimation device 2 is, for example, a GPS receiver. A GPS receiver detects positional information of the own vehicle on the ground by receiving radio waves from artificial satellites. The position information of the own vehicle detected by the GPS receiver includes latitude information and longitude information. The GPS receiver outputs the detected position information of the own vehicle to the controller 100 . Note that the method of detecting the position information of the own vehicle is not limited to the GPS receiver. For example, the position of the ego vehicle may be estimated using a method called odometry. Odometry is a method of estimating the position of the own vehicle by obtaining the movement amount and the movement direction of the own vehicle according to the rotation angle and rotation angular velocity of the own vehicle. A GNSS receiver may be used instead of the GPS receiver.

The controller 100 is a general-purpose microcomputer having a CPU (Central Processing Unit), a memory, and an input/output unit. A computer program is installed in the microcomputer to function as a pedestrian crossing prediction device. By executing the computer program, the microcomputer functions as a plurality of information processing circuits included in the pedestrian crossing prediction device. Here, an example of realizing a plurality of information processing circuits provided in the pedestrian crossing prediction device by software is shown. It is also possible to configure Also, a plurality of information processing circuits may be configured by individual hardware. The controller 100 includes a detection integration unit 4, an object tracking unit 5, an intra-map position estimation unit 6, an action prediction unit 10, and a vehicle control unit 20 as an example of a plurality of information processing circuits. Further, the behavior prediction unit 10 is classified into a pedestrian detection unit 11, a behavior change detection unit 12, a line of sight detection unit 13, an attribute detection unit 14, a statistical data acquisition unit 15, and a likelihood calculation unit 16. .

The detection integration unit 4 integrates a plurality of detection results obtained from each of the plurality of object detection sensors included in the object detection device 1, and outputs one detection result for each object. Specifically, from the behavior of the object obtained from each of the object detection sensors, the most rational behavior of the object with the least error is calculated after considering the error characteristics of each of the object detection sensors. Specifically, by using well-known sensor fusion technology, detection results obtained by a plurality of types of sensors are comprehensively evaluated to obtain more accurate detection results.

An object tracking unit 5 tracks objects detected by the detection integration unit 4 . Specifically, the object tracking unit 5 verifies (associates) the identity of the object between different times based on the behavior of the object output at different times, and tracks the object based on the correspondence.

The in-map position estimation unit 6 estimates the position of the vehicle on the map from the absolute position of the vehicle obtained by the vehicle position estimation device 2 and the map information obtained from the map database 3 . The map database 3 is a database stored in a car navigation device or the like, and stores map information necessary for route guidance such as road information and facility information. Road information is, for example, information relating to intersections, the number of lanes on a road, road boundaries, lane connections, and the like. The map database 3 outputs map information to the on-map position estimation unit 6 in response to a request from the on-map position estimation unit 6 . In this embodiment, the apparatus for predicting pedestrian crossing will be described as having the map database 3 , but the apparatus for predicting pedestrian crossing does not necessarily have the map database 3 . The map information may be acquired by various sensors such as a camera, or may be acquired using vehicle-to-vehicle communication or road-to-vehicle communication. Further, when the map information is stored in a server installed outside, the pedestrian crossing prediction device may acquire the map information from the server at any time through communication. Also, the pedestrian crossing prediction device may periodically obtain the latest map information from the server and update the held map information. The map information stored in the map database 3 may be high definition map data (HD MAP) or normal map data (SD MAP).

Next, details of the function of the behavior prediction unit 10 will be described with reference to FIGS. 2 to 7. FIG. The scene shown in FIG. 2 is a scene in which the own vehicle 50 is traveling toward an intersection. Reference numerals 60-61 indicate pedestrians.

As shown in FIG. 2, when a pedestrian 60 is on one side of a crosswalk and a pedestrian 61 is on the other side, the pedestrians 60 and 61 are detected by the pedestrian detection unit 11. (Step S101 in FIG. 5). More specifically, the pedestrian detection unit 11 detects pedestrians 60 and 61 using data acquired from the object tracking unit 5 . When pedestrians 60 and 61 are detected on both sides of the crosswalk (YES in steps S101 and 103 in FIG. 5), pedestrian detection unit 11 outputs information related to the detection to behavior change detection unit 12. . In FIG. 2, it is assumed that a pedestrian 60 is detected on the left side of the crosswalk and a pedestrian 61 is detected on the right side of the crosswalk. Pedestrian 61 existing on the opposite side of the pedestrian crossing in the direction in which pedestrian 60 travels (the right direction in FIG. 2) may be detected. Moreover, since the detection range is set for the sensors included in the object detection device 1 , the pedestrians 60 and 61 are assumed to be detected within a predetermined distance in front of the vehicle 50 .

The behavior change detection unit 12 detects behavior changes of the pedestrians 60 and 61 detected by the pedestrian detection unit 11 . In this embodiment, the "pedestrian behavior" includes the pedestrian's speed, gesture, hand gesture, face orientation, and body orientation. Therefore, "a change in pedestrian behavior" means a change in the walking speed of the pedestrian, a gesture, a hand wave, a change in the direction of the face, and a change in the direction of the body. In the example of FIG. 2, it is assumed that a change in behavior is first detected in which the pedestrian 60 waves his hand (step S105 in FIG. 5). The behavior change detection unit 12 determines whether or not a behavior change of the pedestrian 61 is detected within a predetermined time after the hand gesture of the pedestrian 60 is detected. The "predetermined time" here is two seconds, for example. Here, it is assumed that a hand gesture (behavior change) of the pedestrian 61 is detected similarly to the hand gesture of the pedestrian 60 within a predetermined time (YES in step S107 of FIG. 5). In this case, the likelihood of the pedestrian 61 crossing the pedestrian crossing is calculated to be higher than when the hand gesture of the pedestrian 61 is not detected. The fact that the hand gesture of the pedestrian 61 is detected within a predetermined time from the hand gesture of the pedestrian 60 means that the pedestrians 60 and 61 are likely to be friends or acquaintances, and that the pedestrian 61 is a pedestrian. This is because it is estimated that there is a high possibility of crossing the pedestrian crossing in order to join 60. In this embodiment, "likelihood" means likelihood and has the concept of high and low. For example, "the pedestrian 61 is likely to cross the pedestrian crossing" means "the pedestrian 61 is likely to cross the pedestrian crossing". The likelihood is calculated by the likelihood calculator 16 . It should be noted that "calculated to have a high likelihood" and "calculated to have a low likelihood" mean, as an example, compared with the case where there is no event that causes the event. The behavior change detection unit 12 outputs information related to detection to the line-of-sight detection unit 13 . An example of calculation of "likelihood of pedestrian 61 crossing the pedestrian crossing" will be mainly described below, but "likelihood of pedestrian 60 crossing the pedestrian crossing" may be calculated in a similar manner. In the present embodiment, it is sufficient to calculate at least one of the "likelihood of the pedestrian 60 crossing the pedestrian crossing" and the "likelihood of the pedestrian 61 crossing the crosswalk". Of course, both likelihoods may be calculated.

When a change in behavior of one pedestrian is detected and a change in behavior of the other pedestrian is detected within a predetermined period of time, the line-of-sight detection unit 13 detects both before and after the timing at which the change in behavior of the pedestrian 61 occurs. pedestrian line of sight. 2, the line-of-sight detection unit 13 detects the line of sight of the pedestrians 60 and 61 because the hand gesture of the pedestrian 61 is detected within a predetermined time after the hand gesture of the pedestrian 60 is detected. (step S109 in FIG. 5). The line-of-sight detection unit 13 detects the lines of sight of the pedestrians 60 and 61 using the information detected by the object detection device 1 . Since a method using a camera image or the like is known for detecting the line of sight, a detailed description thereof will be omitted. Alternatively, line-of-sight detection may be estimated based on the orientation of the faces and the orientation of the bodies of the pedestrians 60 and 61 .

The line-of-sight detection unit 13 determines whether the lines of sight of the pedestrians 60 and 61 match based on the detection result of the lines of sight of the pedestrians 60 and 61 (step S111 in FIG. 5). The likelihood calculation unit 16 changes the likelihood that the pedestrian 61 crosses the pedestrian crossing depending on whether the pedestrians 60 and 61 are in line of sight. When it is determined that the pedestrians' 60 and 61 do not meet the line of sight, the likelihood calculation unit 16 lowers or restores the increased likelihood. It is presumed that the hand gesture of the pedestrian 61 was unrelated to the hand gesture of the pedestrian 60 because "the pedestrians 60 and 61 did not meet their gazes". As a result, the pedestrians 60 and 61 are presumed not to be friends or acquaintances, so it can be said that the possibility of the pedestrian 61 crossing the pedestrian crossing to join the pedestrian 60 is low. Therefore, when it is determined that the lines of sight of the pedestrians 60 and 61 do not meet, the likelihood calculation unit 16 lowers or restores the increased likelihood.

On the other hand, when the line-of-sight detection unit 13 determines that the lines of sight of the pedestrians 60 and 61 are aligned, the likelihood calculation unit 16 changes the likelihood of the pedestrian 61 crossing the pedestrian crossing based on the speed change of the pedestrian 61. . For example, when it is detected that the pedestrian 61 has accelerated, the likelihood calculation unit 16 may increase the likelihood that the pedestrian 61 crosses the pedestrian crossing compared to before the acceleration. Further, when it is detected that the pedestrian 61 decelerates, the likelihood calculation unit 16 may lower the likelihood of the pedestrian 61 crossing the pedestrian crossing compared to before deceleration. Note that the likelihood calculation unit 16 may set the likelihood of one of the pedestrian 60 and the pedestrian 61 that acceleration related to walking is detected to be higher than the likelihood of the other that acceleration is not detected. Further, the likelihood calculation unit 16 may set the likelihood of one of the pedestrian 60 and the pedestrian 61 that deceleration related to walking is detected to be higher than the likelihood of the other that deceleration is not detected.

If the line of sight detection unit 13 determines that the lines of sight of the pedestrians 60 and 61 match (YES in step S111 of FIG. 5), the process proceeds to step S113 of FIG. The attribute detection unit 14 acquires attributes of the pedestrians 60 and 61 . In this embodiment, "pedestrian attributes" include physical characteristics such as height, body width, adult, child, and old age. The attribute detection unit 14 acquires attributes of the pedestrians 60 and 61 using information detected by the object detection device 1 . Pedestrian 60 crosses the pedestrian crossing the likelihood that the pedestrian 61 crosses the pedestrian crossing when the attribute of the pedestrian 60 is detected as "adult" and the attribute of the pedestrian 61 is detected as "child" It may be higher than the likelihood.

When acquisition of the attribute is completed, the process proceeds to step S115 in FIG. The statistical data acquisition unit 15 acquires statistical data. In the present embodiment, "statistical data" is data indicating in which direction the number of people crossing the pedestrian crossing is greater depending on the time of day. For example, consider the case where a station 70 exists at the position shown in FIG. In this case, many pedestrians cross the pedestrian crossing in order to head for the station 70 in the morning hours. That is, as shown in FIG. 3, many pedestrians cross the pedestrian crossing from left to right. On the other hand, in the evening hours, many pedestrians leave the station 70, contrary to the morning hours, so they cross the crosswalk from right to left as shown in FIG. "Statistical data" in the present embodiment is thus statistical data relating to the traveling direction of pedestrians for each time period. The statistical data acquisition method is not particularly limited, but if the statistical data is pre-stored in a storage device (not shown), the statistical data acquisition unit 15 can acquire the statistical data by referring to the storage device. Alternatively, statistical data may be obtained by road-to-vehicle communication. The likelihood calculation unit 16 may calculate the likelihood using statistical data. For example, in the morning time period, the likelihood calculation unit 16 may increase the likelihood that the pedestrian 60 crosses the pedestrian crossing more than the pedestrian 61 crosses the pedestrian crossing. On the other hand, the likelihood calculation unit 16 may increase the likelihood of the pedestrian 61 crossing the pedestrian crossing over the likelihood of the pedestrian 60 crossing the pedestrian crossing in the night time zone.

The likelihood calculation unit 16 calculates the likelihood that the pedestrians 60 and 61 will cross the pedestrian crossing based on behavior change, line of sight, walking speed, pedestrian attributes, statistical data, and the like (step S117 in FIG. 5). As described above, the likelihood calculation unit 16 changes the likelihood that the pedestrians 60 and 61 cross the pedestrian crossing based on each detected element, and adds the likelihood calculated based on each element. If the total likelihood value exceeds a predetermined value (eg, 0.7), the likelihood calculation unit 16 predicts that the target pedestrian among the pedestrians 60 and 61 will cross the pedestrian crossing. The likelihood calculator 16 outputs the predicted result to the vehicle controller 20 .

When the likelihood calculation unit 16 predicts that the target pedestrian among the pedestrians 60 and 61 will cross the pedestrian crossing, the vehicle control unit 20 accelerates the deceleration of the own vehicle 50 so that the vehicle 50 reaches the crosswalk before the pedestrian crossing. Vehicle 50 is stopped (step S119 in FIG. 5). As a result, for example, when the host vehicle 50 is a vehicle equipped with an automatic driving function, the sense of discomfort felt by the occupants regarding automatic driving is eliminated or reduced.

(Effect)
As described above, the pedestrian crossing prediction device according to the present embodiment provides the following effects.

The pedestrian crossing prediction device includes a sensor (object detection device 1 ) mounted on own vehicle 50 and controller 100 . The sensor detects a pedestrian 60 (first pedestrian) within a predetermined distance in front of the vehicle 50 . The sensor detects a pedestrian 61 (second pedestrian) on the other side of the road with the position of the pedestrian 60 as a reference. The controller 100 compares the case where the behavior change of the pedestrian 61 is detected within a predetermined time after the sensor detects the behavior change of the pedestrian 60 to the case where the behavior change of the pedestrian 61 is not detected within the predetermined time. At least one of the first likelihood that the pedestrian 60 crosses the road and the second likelihood that the pedestrian 61 crosses the road is increased. As a result, when there are pedestrians 60 and 61 on both sides of a pedestrian crossing as shown in FIG. 2, it is possible to accurately predict whether at least one of the pedestrians will cross the pedestrian crossing. Become.

If a change in behavior of pedestrian 61 (second pedestrian) is detected within a predetermined time after the change in behavior of pedestrian 60 (first pedestrian) is detected by the sensor, controller 100 detects the change in behavior of pedestrian 61 (second pedestrian). It is determined whether or not the line of sight of the pedestrian 60 and the pedestrian 61 is met before and after the timing when the behavior change occurs. The controller 100 changes at least one of the first likelihood and the second likelihood based on the determination result of whether or not the line of sight is met. By determining whether or not the line of sight meets at such timing, it is possible to estimate whether or not the pedestrians 60 and 61 are friends or acquaintances. Therefore, by changing the likelihood using the determination result of whether or not the lines of sight meet, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

The controller 100 may lower the increased likelihood when it is determined that the line of sight does not match. The fact that the pedestrians' 60 and 61 did not meet their gazes is presumed that the hand gesture of the pedestrian 61 was irrelevant to the hand gesture of the pedestrian 60 . As a result, the pedestrians 60 and 61 are presumed not to be friends or acquaintances, so it can be said that the possibility of the pedestrian 61 crossing the pedestrian crossing to join the pedestrian 60 is low. Therefore, when it is determined that the lines of sight of the pedestrians 60 and 61 do not meet, the controller 100 may lower the increased likelihood. This makes it possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

When the controller 100 determines that the line of sight is met, the controller 100 selects one of the first likelihood and the second likelihood based on the change in walking speed of the pedestrian 60 (first pedestrian) and the pedestrian 61 (second pedestrian). , at least one likelihood may be changed. By changing the likelihood based on the speed change related to walking, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

The controller 100 makes the likelihood of one of the pedestrian 60 (first pedestrian) and the pedestrian 61 (second pedestrian) that acceleration related to walking is detected higher than the likelihood of the other that acceleration related to walking is not detected. may For example, if pedestrian 61 is detected to have accelerated but pedestrian 60 has not been detected to accelerate, controller 100 determines the likelihood that pedestrian 61 will cross the pedestrian crossing. It may be higher than the likelihood. It is estimated that the pedestrian 61 is likely to cross the pedestrian crossing in order to join the pedestrian 60 when the acceleration related to walking is detected. Therefore, by increasing the likelihood according to acceleration, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

The controller 100 lowers the likelihood of one of the pedestrian 60 (first pedestrian) and pedestrian 61 (second pedestrian) that deceleration related to walking has been detected than the likelihood of the other that deceleration has not been detected. may For example, if pedestrian 61 is detected to have slowed down while pedestrian 60 is not detected to have slowed down, controller 100 determines the likelihood that pedestrian 61 will cross the pedestrian crossing. It may be lower than the likelihood. It is presumed that the pedestrian 61 has given up crossing the crosswalk because deceleration related to walking has been detected. Therefore, by lowering the likelihood according to deceleration, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

The controller 100 compares the likelihood of one of the pedestrian 60 (first pedestrian) and pedestrian 61 (second pedestrian) whose attribute is detected as a child to the likelihood of the other whose attribute is not detected as a child. You can make it higher. For example, when the attribute of pedestrian 61 is detected as child and the attribute of pedestrian 60 is detected as adult, controller 100 determines the likelihood that pedestrian 61 will cross the pedestrian crossing. may be higher than This makes it possible to take into account the instability of the child's behavior.

The controller 100 may acquire statistical data regarding the traveling direction of pedestrians for each time period. The controller 100 may change at least one of the first likelihood and the second likelihood based on the obtained statistical data. For example, as shown in FIG. 3, controller 100 may increase the likelihood of pedestrian 60 crossing the pedestrian crossing over the likelihood of pedestrian 61 crossing the pedestrian crossing during the morning hours. On the other hand, as shown in FIG. 4, if it is nighttime, the controller 100 may make the likelihood of the pedestrian 61 crossing the pedestrian crossing higher than the likelihood of the pedestrian 60 crossing the pedestrian crossing. This makes it possible to accurately predict whether or not at least one of the pedestrians will cross the pedestrian crossing.

The controller 100 controls the pedestrian 60 (first pedestrian) and the pedestrian 61 (second pedestrian) in a case where the traveling direction of the pedestrian 60 (first pedestrian) and the pedestrian 61 (second pedestrian) is parallel to the traveling direction of the road, and the pedestrian 60 and the pedestrian 61, If either deceleration or stoppage is detected, or if no change in behavior of the pedestrian 61 is detected within a predetermined time, at least one of the first likelihood and the second likelihood is Compared to the case where deceleration or stoppage of either the person 60 or the pedestrian 61 is not detected, or the case where the behavioral change of the pedestrian 61 is detected within the predetermined time, it may be lowered. Consider a case where pedestrians 60 and 61 are parallel to the direction of travel on the road, as shown in FIG. In this scene, if either one of the pedestrians 60 and 61 decelerates or stops, or if no change in the behavior of the pedestrian 61 is detected within a predetermined period of time, the acquaintance across the road It is presumed that there is a high possibility that they met by chance and exchanged brief greetings. It is estimated that both (or at least one of) the pedestrian 60 and the pedestrian 61 are less likely to bother to cross the road to meet each other. Therefore, by lowering the likelihood, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the road. Note that if a change in the behavior of the pedestrian 61 is not detected within a predetermined period of time after the deceleration or stoppage of one of the pedestrian 60 and the pedestrian 61 is detected, the controller 100 may lower the likelihood. good.

The controller 100 controls the direction of travel of the pedestrian 60 (first pedestrian) and the pedestrian 61 (second pedestrian) in parallel with the running direction of the road, and the direction of travel is in the direction of the crosswalk. In the case of a change toward the direction of travel, the likelihood that one direction of travel has changed may be higher than the other likelihood that the direction of travel has not changed. For example, as shown in FIG. 7, when the pedestrian 60 is detected, the traveling direction of the pedestrian 60 is parallel to the running direction of the road. There are cases where the direction of travel changes toward the direction. This change indicates that the pedestrian 60 is likely to cross the pedestrian crossing. Therefore, by calculating the likelihood based on this change, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the road.

Behavior changes include speed changes, gestures, hand gestures, and face orientations related to walking. By calculating the likelihood using such behavior changes, it becomes possible to accurately predict whether or not at least one of the pedestrians will cross the road.

When at least one of the first likelihood and the second likelihood is greater than a predetermined value, the controller 100 accelerates the deceleration of the own vehicle 50 and stops it before the crosswalk. As a result, for example, when the host vehicle 50 is a vehicle equipped with an automatic driving function, the sense of discomfort felt by the occupants regarding automatic driving is eliminated or reduced.

Each function described in the above embodiments may be implemented by one or more processing circuits. Processing circuitry includes programmed processing devices, such as processing devices that include electrical circuitry. Processing circuitry also includes devices such as application specific integrated circuits (ASICs) and circuit components arranged to perform the described functions.

While embodiments of the present invention have been described above, the discussion and drawings forming part of this disclosure should not be construed as limiting the invention. Various alternative embodiments, implementations and operational techniques will become apparent to those skilled in the art from this disclosure.

In the above-described embodiment, the case where the pedestrians 60 and 61 are detected across the pedestrian crossing has been described, but this is not a limitation. is also applicable for cases detected on both sides of

1 object detection device, 2 vehicle position estimation device, 3 map database, 4 detection integration unit, 5 object tracking unit, 6 map position estimation unit, 10 behavior prediction unit, 11 pedestrian detection unit, 12 behavior change detection unit, 13 line-of-sight detection unit 14 attribute detection unit 15 statistical data acquisition unit 16 likelihood calculation unit 20 vehicle control unit 100 controller

Claims (13)

A pedestrian crossing prediction method for a pedestrian crossing prediction device equipped with a sensor and a controller mounted on a vehicle,
The sensor is
detecting a first pedestrian within a predetermined distance in front of the own vehicle;
Detecting a second pedestrian on the other side of the road on the basis of the position of the first pedestrian,
The controller is
When the change in behavior of the second pedestrian is detected within a predetermined time after the change in behavior of the first pedestrian is detected by the sensor, the change in behavior of the second pedestrian is detected within the predetermined time. At least one of the first likelihood that the first pedestrian crosses the road and the second likelihood that the second pedestrian crosses the road is increased compared to the case where the Pedestrian crossing prediction method. The controller is
When the change in behavior of the second pedestrian is detected within a predetermined time after the change in behavior of the first pedestrian is detected by the sensor, before and after the timing at which the change in behavior of the second pedestrian occurs Determining whether the first pedestrian and the second pedestrian's line of sight match,
2. The pedestrian crossing prediction according to claim 1, wherein at least one of the first likelihood and the second likelihood is changed based on the determination result of whether or not the line of sight is met. Method. The controller is
3. The pedestrian crossing prediction method according to claim 2, wherein when it is determined that the lines of sight do not meet, the increased likelihood is decreased. The controller is
When it is determined that the lines of sight match, at least one of the first likelihood and the second likelihood is changed based on a change in walking speed of the first pedestrian and the second pedestrian. The pedestrian crossing prediction method according to claim 2, characterized in that: The controller makes the likelihood of one of the first pedestrian and the second pedestrian that acceleration related to walking is detected higher than the likelihood of the other that the acceleration is not detected. Item 5. The pedestrian crossing prediction method according to item 4. The controller makes the likelihood of one of the first pedestrian and the second pedestrian that deceleration related to walking is detected lower than the likelihood of the other that the deceleration is not detected. Item 5. The pedestrian crossing prediction method according to item 4. The controller makes the likelihood of one of the first pedestrian and the second pedestrian whose attribute is detected as child higher than the likelihood of the other pedestrian whose attribute is not detected as child. A pedestrian crossing prediction method according to any one of claims 1 to 6. The controller is
Get statistical data on the direction of travel of pedestrians by time of day,
The pedestrian according to any one of claims 1 to 7, wherein at least one of the first likelihood and the second likelihood is changed based on the acquired statistical data. Traverse prediction method. The controller controls any one of the first pedestrian and the second pedestrian when the traveling directions of the first pedestrian and the second pedestrian are parallel to the traveling direction of the road. If the deceleration or stop of is detected, or if the behavior change of the second pedestrian is not detected within the predetermined time, at least one likelihood of the first likelihood and the second likelihood , when either one of the first pedestrian and the second pedestrian is not detected to decelerate or stop, or when a change in behavior of the second pedestrian is detected within the predetermined time. 9. The pedestrian crossing prediction method according to any one of claims 1 to 8, wherein the pedestrian crossing prediction method according to any one of claims 1 to 8, wherein the height is lowered. When the direction of travel of the first pedestrian and the second pedestrian is parallel to the running direction of the road, and the direction of travel changes toward the direction in which the crosswalk exists, The pedestrian crossing prediction method according to any one of claims 1 to 9, characterized in that the likelihood of one of said moving directions changing is set higher than the other likelihood of said moving directions not changing. The pedestrian crossing prediction method according to any one of claims 1 to 10, wherein the behavioral changes include speed changes, gestures, hand gestures, and face orientations related to walking. 12. The controller according to any one of claims 1 to 11, wherein when at least one of the first likelihood and the second likelihood is greater than a predetermined value, the controller accelerates the deceleration of the host vehicle. The pedestrian crossing prediction method according to item 1. A pedestrian crossing prediction device comprising a sensor and a controller mounted on the own vehicle,
The sensor is
detecting a first pedestrian within a predetermined distance in front of the own vehicle;
Detecting a second pedestrian on the other side of the road on the basis of the position of the first pedestrian,
The controller is
When the change in behavior of the second pedestrian is detected within a predetermined time after the change in behavior of the first pedestrian is detected by the sensor, the change in behavior of the second pedestrian is detected within the predetermined time. At least one of the first likelihood that the first pedestrian crosses the road and the second likelihood that the second pedestrian crosses the road is increased compared to the case where the Pedestrian crossing prediction device.

Images